Stokes drift induced by ocean surface waves is a key component of wave–current interactions and plays an important role in upper-ocean transport processes. Despite its significance, direct observations of Stokes drift remain challenging, and estimates are generally derived from wave spectra measurements. In situ wave observations from wave buoys provide reliable spectral information but are spatially sparse and primarily available in limited regions. Satellite remote sensing offers the opportunity for global wave spectrum observations, with Synthetic Aperture Radar (SAR) and the Surface Wave Investigation and Monitoring (SWIM) mission providing complementary capabilities. SAR can resolve wave directions but is affected by severe spectral distortions. In contrast, SWIM provides extensive spectral coverage with relatively small amplitude bias but suffers from directional ambiguities. Here, we develop a cross-sensor wave spectrum fusion method that integrates SAR and SWIM observations to overcome their individual limitations and improve Stokes drift estimation. The fused Stokes drift products are evaluated against measurements from wave buoys deployed at Ocean Station Papa, Southeast Hawaii, and the Southern Ocean Flux Station. The results show that the fusion method consistently reduces uncertainties compared with estimates derived independently from SAR or SWIM, with improved agreement of the mean Stokes drift vector across different sea states. We further analyze the error characteristics of the fusion framework and identify several dominant uncertainty sources that provide insights for future refinement. Finally, global application of the fused product reveals seasonal Stokes drift distributions that better reproduce known global wave climatological patterns than SWIM-only estimates. This work demonstrates the potential of satellite wave spectrum fusion for advancing global-scale estimates of wave-driven ocean processes.